Osteoporos Int (2002) 13:765–767
ß 2002 International Osteoporosis Foundation and National Osteoporosis Foundation
Economic Evaluation of Interventions for Osteoporosis
J. A. Kanis1 and B. Jonsson2
Centre for Metabolic Bone Diseases (WHO Collaborating Centre), University of Shefﬁeld Medical School, Shefﬁeld, UK;
Department of Economics, Stockholm School of Economics, Stockholm, Sweden
Health economic evaluation is an increasingly important model. It is suggested that the model developed at the
component of drug development in osteoporosis because Stockholm School of Economics  be used as a
of the rapidly-increasing opportunities for prevention reference standard. Different, and usually more compli-
and treatment. There is a need to justify resource cated approaches are under development, but the
allocation both for the selection of therapies for advantages of these have yet to be shown. A key
prevention and treatment of osteoporosis, and for factor is the availablity of data to populate more
allocating resources for this condition in competition sophisticated models. A reference standard is, however,
with other diseases. The ﬁeld of osteoporosis is only a platform from which new approaches can be
immature in terms of pharmaceutical development judged in the search for improvements.
compared to many established chronic diseases such as With regard to the selection of the primary effective-
cardiovascular disease or diabetes despite the substantial ness or outcome measure for a model, fracture outcomes
health economic burden of osteoporotic fractures . are preferred to bone mineral density. Whereas there is a
Treatment and prevention strategies for osteoporosis well-established relationship between bone mineral
must therefore be shown to be not only medically density (BMD) and fracture risk, many factors other
justiﬁed and possible to implement in clinical practice, than BMD contribute to risk. Moreover, the treatment-
but also cost-effective. Model-based health economic induced changes in fracture rate cannot be inferred from
studies provide a cornerstone because of their opportu- the magnitude of treatment-induced changes in BMD
nity to integrate epidemiological, clinical and economic with any degree of conﬁdence in that changes in fracture
data and their ﬂexibility to take account of the many rate are greater than that expected solely from changes in
uncertainties that confront such analyses. BMD . Where possible, reduction in fracture rates and
This supplement issue provides an update for the other outcomes should be modeled from information
health economic assessment of interventions in osteo- available from randomized controlled trials (RCTs).
porosis. It builds on an earlier platform published several All relevant outcomes should be considered and
years previously . The ﬁrst paper, developed under the modeled separately. This increases the complexity of
auspices of the two WHO collaborating centers in the modeling due to the increased number of transition
ﬁeld, provides an extensive review of the data available states, but the complexity generated by the great number
to date on cost utility analysis. The preferred approach is of potential osteoporotic fractures, can for some
cost utility analysis, involving the calculation of cost per applications be reduced by weighting fracture incidence
quality adjusted life year gained from intervention, since by disutility . For several treatments in osteoporosis,
such analyses provide an established mechanism for extraskeletal beneﬁts and risks should be accommodated.
comparing cost effectiveness across diseases – an Examples include the effects of estrogens and selective
important component of investment strategies. The estrogen receptor modulators on cardiovascular disease
preferred approach to modeling is the use of a Markov and breast cancer. For agents without signiﬁcant
extraskeletal beneﬁts or risks this complexity is
Correspondence and offprint requests to: Prof Kanis, WHO redundant. Notwithstanding, drugs do have side effects
Collaborating Centre for Metabolic Bone Diseases, University of with very signiﬁcant implications for cost utility, and
Shefﬁeld Medical School, Beech Hill Road, Shefﬁeld S10 2RX, UK. models should provide for this contingency.
766 J. A. Kanis and B. Jonsson
It should be acknowledged that our information on attributed both to hip fracture and spine fracture
fracture risk is incomplete and all assumptions should be (though not for Colles fracture). A great uncertainty is
made explicit. For example, in modeling cohorts of the extent to which mortality is reversible by interven-
patients with osteoporosis but without fracture, the tion and few studies have yet addressed this point.
relevant fracture risk is the risk of a ﬁrst fracture. The greatest difﬁculty that arises in making recom-
Crude data on incidence overestimate this risk . In mendations is the perspective of cost-effective analysis.
patients with established osteoporosis the risk of future On the one hand it is argued that cost-effectiveness
fracture may not be linear. In individuals hospitalized for should include all costs and beneﬁts to society – the
vertebral fracture the risk of further fracture is increased societal perspective. On the other hand, healthcare
markedly in the several months following admission and purchasers, regulatory agencies and reimbursement
then tails off, though not to the risk of the general committees are often only interested in costs that affect
population . The risk of fracture also varies markedly their budget. A decision maker may also only be
in different countries so that appropriate data should be interested in costs and beneﬁts over a limited time
used or the assumptions made explicit. period suitable for their own decisions. A decision maker
Emphasis is also placed on several aspects of may thus decide to ignore costs due to reduced mortality
treatment. Treatment effects should be modeled in a (cost in added years of life) despite the arguments for
base case from randomized controlled trials where including these costs in a social perspective that are
available with appropriate assumptions on how long presented in the report. Thus, the recommendation arises
the intervention lasts, compliance and the offset of to preserve ﬂexibility intrinsic in the model and to
effects once treatment is terminated. It is appropriate to suggest that, where feasible, cost utility analysis should
state the levels of evidence on which efﬁcacy is assumed also include an assessment of the societal perspective.
in line with recommendations for evidence-based There will, for example, be instances where decisions
medicine. This would help avoid the entrenched favorable for a purchaser are unfavorable for society and
positions of dogma, for example the view that hormone vice versa, and that information may be valuable for
replacement provides signiﬁcant cardiovascular protec- understanding the incentives and disincentives for
tion – a view that is being challenged as data from prevention and treatment of osteoporosis that different
randomized controlled trials become available . decision makers face.
However, the request for evidence-based data on A second paper developed by the International
effectiveness should not prevent the use of sensitivity Osteoporosis Foundation reviews the information on
analysis to address ‘what if’ issues where uncertainty of utilities available in osteoporosis. This highlights several
effectiveness is genuine. uncertainties. An example is provided by the health
A crucial issue concerning effectiveness of interven- states assigned to the same osteoporotic fractures by
tion relates to what happens when treatment is stopped. different methodologies (e.g. standard gamble, time
Fracture risk is unlikely to revert to that of the untreated trade off). Differences also arise depending on whether
patient at the time of stopping treatment. Rather, the utility values are derived from patients, members of
effect will wane over several years. Economic analysis society or expert committees. There is also a lack of
has shown the great importance of assumptions in offset empirical data concerning the time course of disutility
time . Unfortunately, offset times have not been well following many osteoporotic fractures which is clearly
studied, and this lies behind the recommendations by the an area for further research.
WHO for this to be studied as a requisite of drug
The review also makes recommendations for a deﬁned References
standard for cost effectiveness. A value of $60 000 per
quality of life year (QALY) gained is suggested for the 1. World Health Organization. Assessment of fracture risk and its
application to screening for postmenopausal osteoporosis. WHO
developed countries. It is relevant to note that this is Technical Report Series 843. WHO: Geneva, 1994.
based on analyses that include the cost of added years 2. World Health Organization. Recommendations for health
, and when these are excluded a lower value of economic evaluations of interventions in osteoporosis (WHO/
$30 000 per QALY gained provides a reasonable yard- NCD/NCM/CRA/99.3). WHO: Geneva, 1999.
3. Zethraeus N, Johannesson M, Jonsson B. A computer model to
stick. The reason for the higher value when costs of added analyse the cost-effectiveness of hormone replacement therapy.
years are included is that cost-effectiveness ratios Int J Technol Assess Health Care 1999;15:352–65.
generally increase when the costs of added life years 4. Black DM, Pearson J, Harris F, La Croix A, Cummings SR.
are included, since most reductions in mortality occur at Predicting the effect of anti-resorptive treatments on risk of
older ages. The inclusion of costs in added years of life vertebral fractures: a meta-analysis. J Bone Miner Res 1999;14
will also change the relative cost-effectiveness of differ- 5. Kanis JA, Oden A, Johnell O, Jonsson B, De Laet C, Dawson A.
ent programs. Mortality reductions at younger ages and The burden of osteoporotic fractures: a method for setting
QALY gains at older ages will affect cost-effectiveness intervention thresholds. Osteoporos Int 2001;12:417–27.
of interventions aimed at the young or the elderly. 6. Kanis JA, Johnell O, Oden A, et al. Long-term risk of
osteoporotic fracture in Malmo. Osteoporos Int 2000;11:669–674.
It is important to acknowledge the incompleteness of 7. Johnell O, Oden A, Caulin F, Kanis JA. Acute and long-term
information. An example is provided by the estimates of increase in fracture risk after hospitalization for vertebral fracture.
mortality following fracture. Excess mortality is Osteoporos Int 2001;12:207–14.
Economic Evaluation of Interventions for Osteopoosis 767
8. Hulley S, Grady D, Bush T, et al. for the Heart and Estrogen/ offset of effect of treatments for hip fracture on health outcomes.
Progestin Replacement Study (HERS) Research Group. Rando- Osteoporosis Int 1999;10;193–9.
mized trial of estrogen plus progestin for secondary prevention of 10. World Health Organization. Guidelines for preclinical evaluation
coronary heart disease in postmenopausal women. JAMA and clinical trials in osteoporosis. WHO: Geneva, 1998.
1998;280:605–13. 11. Meltzer D. Accounting for future costs in medical cost-
9. Jonsson B, Kanis JA, Dawson A, Oden A, Johnell O. Effect and effectiveness analysis. J Health Econ 1997;16:33–64.
Received for publication 6 May 2002
Accepted 7 May 2002